Agriculture Reference
In-Depth Information
Where
TME
= true ME content of feed (kcal/g);
CP
= crude protein (g/
100
g); and
Q
is a cali-
bration coefficient of default value =
1.
The
units of the number 5.739 are kcal/g protein.
From the bird data two polynomial
equations were calculated for live weight
(
LW F
) more or less than 640 g. For the small-
er birds the relationship derived was:
RU
(per bird day) = -7.110241 + 2.549947
(
LW F
) - 4.267 ×
10
-3
(
LW F
)
2
+ 4.0 ×
10
-6
(
LW F
)
3
(
n
=
11,
R
2
= 1.00)
in a meaningful way. A device called the
'bilan thermique' (thermostat) adjusts feed
intake relative to the reference level follow-
ing calculation of the energy balance. The
reference data are also used to adjust two
parameters in the model prior to modelling;
this adjustment makes the reference simula-
tion and the reference data coincide and,
the authors claim, allows the model to oper-
ate over the very wide range of environ-
ments and differences in bird activity re-
flected in the French production sectors.
The parameters adjusted are called 'l'indice
d'entretien' (the index of maintenance) and
the 'facteur d'activité' (factor of activity)
and the adjusted values are used in the sub-
sequent simulations. The model then con-
siders inputs for diet composition, feed
form and pellet quality, physical activity
and effective temperature as influenced by
ambient temperature, humidity, air speed,
stocking density and sun radiation. Both
constant and cyclic temperatures are con-
sidered and a lighting module distinguishes
between light and dark periods. The model
cycles on a period of
1
h of the bird's life.
The look-up facilities of Vensim
®
offer
the user graphical control of some
19
com-
ponents of the model.
Figure 1.1
illustrates
one of these for adjusting the influence of air
speed on effective temperature in hot condi-
tions. While some of these adjustments are
quite complex they undoubtedly offer a
very valuable facility for use of this model
and for its further development.
The turkey growth model described by
Rivera-Torres
et al
. (2011; see also Rivera-
Torres, Chapter
8,
this volume) is included
in
Table 1.2
even though this work draws on
a different strand of modelling theory and
technique. This started with a general paper
by Sauvant (1992) developed initially for
application to ruminants (Sauvant, 1994 and
elsewhere), but also applied to pigs (Lovatto
and Sauvant, 2003) and in this case to turkeys.
The model works at an 'operational' level,
which is roughly comparable to the models
described above, but also includes a 'regula-
tory' subsystem (called the 'decisional' system
by Rivera-Torres), which is described at a
lower level of metabolic organization within
the animal. The two systems interact so that
(1.5)
Food intake is calculated as
FI
= (
RU
/bird
day)/(
RU
/g diet) ×
F
g/bird day, where
F
is a
calibration coefficient of default value =
1.
Apart from repeating all the observa-
tions and calculations it is difficult to see
how such an approach can accommodate
genetic change or the subtle differences be-
tween broiler strains which are a feature of
the modern industry.
The
two remaining models listed in
Table 1.2
are discussed elsewhere in this
volume but a brief mention is included here
to compare the approaches used. The model
INAVI developed and described by Quentin
(2004; see also Méda
et al
., Chapter
9,
this vol-
ume) represents a significant recent devel-
opment in the field of poultry modelling. The
program was developed using the Vensim
®
modelling software in combination with
Microsoft Excel spreadsheets. Vensim
®
pro-
vides a useful facility for the user to adjust
some of the more difficult functions in the
model using a graphical 'look-up' facility.
This facilitates the use of 'local' data and
also allows for these aspects of the model to
be continuously updated. INAVI is specific-
ally aimed at commercial practice in the
French market and the authors include a
wide range of genotypes and environments
that cover the main systems of production
in France; commercial broiler, Label and the
intermediate 'Certified' production.
INAVI is essentially driven by feed in-
take. For this, and for a range of other model
elements, reference levels of bird perform-
ance are used to initiate the simulation. It is
the user's responsibility to ensure that the
reference data for different inputs are related
